How has the evidence for
stroke care progressed over
the past decade?
10th Anniversary UKSF
2nd December 2015
Peter Sandercock
University of Edinburgh
Huge increase over the decade
in knowledge of stroke and its
treatment
Epidemiology
Acute treatment
Rehabilitation
Prevention
Global burden of stroke: mortality
Lancet Neurology 2009; 8:345–354
10 treatment headlines for the decade
2005 Early supported discharge services
2007 Hemicraniectomy for malignant MCA stroke
2008 Time window for iv rt-PA extends to 4.5 hours
2009 Novel anticoagulants for stroke prevention in AF
2010 GCS ineffective for post-stroke DVT prevention
2012 Statins for prevention stroke & MI
2013 Organised inpatient care (Stroke Unit) update
2014 iv rt-PA for acute ischaemic stroke – who benefits?
2014 IPC for DVT prevention
2015 Mechanical clot retrieval for acute ischaemic stroke
2015 Effects of Very Early Rehabilitation
10 treatment headlines for the decade
2005 Early supported discharge services
2007 Hemicraniectomy for malignant MCA stroke
2008 Time window for iv rt-PA extends to 4.5 hours
2009 Novel anticoagulants for stroke prevention in AF
2010 GCS ineffective for post-stroke DVT prevention
2012 Statins for prevention stroke & MI
2013 Organised inpatient care (Stroke Unit) update
2014 iv rt-PA for acute ischaemic stroke – who benefits?
2014 IPC for DVT prevention
2015 Mechanical clot retrieval for acute ischaemic stroke
2015 Effects of Very Early Rehabilitation
Evidence: past problems• ‘Breakthrough science’ often not replicated1
• 85% of research is wasted,2 usually because it
was
– Not published or poorly reported
– Asked the wrong questions
– Badly designed
– Failed to review existing research systematically
before starting new work
• Huge information overload for us all.
1. Why most published research findings are false. Ioannidis, Plos Med 2005. 2. Research Waste series Lancet 2015
2005-15: Doubling in number of reports of
trials 12,000 –> 24,000
Number of reports of trials
Have research METHODS improved?
• EQUATOR guidelines for how to report research
• Better at asking the right research questions
– James Lind Alliance: Research priority setting
– Measuring the right core outcomes (COMET)
– Stakeholder involvement
• Help to increase research value & reduce waste
– REWARD Alliance researchwaste.net
– Trial Forge; gathering evidence on how to make trials more
efficient www.trialforge.org
• Cochrane Collaboration making continual efforts to
improve quality and relevance of its systematic
reviews with greater focus on priority topics
= 100 new topics reviewed and many important
reviews of priority topics updated
Number of Cochrane stroke
systematic review topics
Evidence is cumulative: effect of stroke unit care
on being ‘dead or needing institutional care’
• The question has been
answered
• There is ‘proof beyond
reasonable doubt’ of the
benefits
• The next question is
‘can this be translated
to low and middle
income countries?’
.3 .5 1 2 5
1960
1970
1980
1990
2000
Stroke unit better
SUTC (unpublished)
Stroke unit worse
Stroke Unit care in low- and middle-income countries reduces case-fatality
Langhorne and Pandian Lancet Neurol 2012
Benefit of stroke unit care
• Independent of parent specialty1
• Demonstrated in all regions
studied, including low-and middle
income countries1
• Not dependent on ‘technology’ or
CT scan
• Result from reduction in stroke
complications2
• Global impact -> WHOStroke Unit Trialists Collaboration Cochrane Library 2007, Pandian Lancet Neurology 2012. Govan et al. Stroke 2007
2015 WHO meeting on managing
stroke and heart disease• Previous WHO focus for reducing burden of non-
communicable disease was prevention, but…
• Systems of care for people with stroke and heart
disease patchy and inadequate
• In low and middle income countries, WHO
considering new initiative to get health systems to
– ‘Do the simple things well’
– Ensure equitable access to basic acute stroke care
facilities for all
Advances in rehabilitation researchNew research designs
• Multicentre Trials
• Methods to develop trials of complex interventions
Discipline-specific
• Speech therapy
• Occupational therapy
• Physiotherapy
New technology
• CIMT
• Transcranial stimulation– Electrical
– Magnetic
• Robotics
Nurse
ICONS
OTCH
Robotics for arm therapy: effect on ADL
Mehrholz et al CDSR 2015
Conclusion: robotics for the
post-stroke arm & hand
Implications for practice
• electromechanical and robot-assisted arm and hand
training after stroke might improve activities of daily
living & arm and hand function…
• it is still not clear if the difference… is clinically
meaningful
Implications for research
• There is still a need for well-designed, large-scale,
multicentre studies to evaluate benefits and harms
Mehrholz et al CDSR 2015
Intervention Maximum impact
Extra independent
survivors if all
patients were
eligible
Proportion eligible
for treatment (%)
Actual impact
Extra
independent
survivors per
year
Rapid secondary prevention 200 All TIA; 15 stroke 100
Stroke unit (CSU) service 500 80 400
Rehabilitation (ESD) service 500 25 130
Service total - - 630
Aspirin 100 80 80
rtPA within 0-3 hrs
rtPA within 3-4.5 hrs
1100
500
10
10
160
Mechanical thrombectomy 1200 10 120
Hemicraniectomy 2000 0.5 10
Acute medical total - - 370
Total - - 1000
Population impact of services in 5 million inhabitants
(10,000 strokes per year): number of extra independent
survivors per year
Table courtesy of Peter Langhorne and Jeyaraj Pandian
How has the evidence ‘progressed’?
• We understand better what the problems are
• We now know much more reliably
– which interventions are effective
– which are definitely not
• We have better methods for research:
– Design
– Conduct, analysis and reporting
– Evidence synthesis
– Priority-setting for and planning of new research
• Conclusion: the past decade of research has clearly
– Improved patient care
– Helped improve future research (and make it less wasteful?)
Thank you
Acknowledgements
Cochrane Stroke Group
Peter Langhorne and Jeyaraj Pandian